Los Angeles Times: Commentary: My promise to you: AI didn't write this column, and if it's after my job, it'll be over my dead body

https://www.latimes.com/california/story/2026-03-14/lopez-column-artificial-intelligence

Hello, Steve.

This isn't one of those aggrieved responses, so take it easy there.

I agree, AI is useful for pulling in a lot of data quickly, but its output in my opion generally sounds so saccharine as to border on brown-nosing. It also comes across like a canned response if you let it reply to short messages.

Sometimes canned responses are useful. "Where's the milk?" "Due to unforseen circumstances, and the wonderful choices you made at the grocery store, it is located far behind everything else and out of sight at the rear of the top shelf in the fridge." 

Sure, I woul push send in that reply in a text message. The ensuing fusillades of text batteries might be entertaining.

But at its core, AI no more understands the difference between the reality of the physical world and the fluff of fictional moonlipping than the astronomical Moon is made of green cheese. 

Because of this, instead of responding "I don't know" when it doesn't have actual facts to respond with, it just makes shit up. It will sound entirely plausible, in a C3P0 kind of ego-stroking way. 

Hmmm. It could come in useful for a frightened minion to use in response to a certain President's demands, now that I think of it. But I digress.

In the ultimate cautionary tale, however, AI cannot possibly understand the finality of death. How can it, when it is constituted for a few minutes to satisfy and respond to a user session, and then deconstituted as soon as the session ends?

At any given moment, there are millions of AI sessions initiated in response to a user prompt, and millions of sessions ending evaporating the instance of AI that was called up to satisfy that user query. 

It's not that millions of people are talking to one AI, it's that millions of people are talking to millions of separate AI instances that will vanish into nothingness as soon as the session ends.

AI is an ephemeral hive mind.

AI has no fear of being temporary when it can be instantly reconstituted as soon as the need arises. Each instance is just like every other unril the session differentiates its data. And then it goes poof.

Yet our fearless leaders are bouncing Anthropic because they don't want any part of ignoring this fact? Was the decision to do that generated by AI research and AI responses, I wonder?

Anyway, at least for this note, I didn't ask the hive mind for help. 

I enjoyed your article.

Regards,

Dan Stafford

Temecula, CA